SOCI 421 Lecture 13: (13) Sociological Statistics
2 Mar 2016
School
Department
Course
Professor
Document Summary
Labelling variables that do not have a quantitative value. Transform into lower-order scales, but not in reverse. E. g. age, level of education, number of crimes committed. Association: do not require to distinguish between x/y, just find a relationship. B= slope, how much y changes with one-unit increase in. B = . 5, then u increases . 5 with one unit increase in x. Tells us about the average relationship between x/y. Purpose: examine independent relationship/outcome in set of preduictors. Allows comparison of relative influence of different predictors. X= structural variables, m=family process variables, y = Ensure x is related to m. m becomes y. X -> m or structural variables -> family process variables. M is now a predictor of our final outcome, y. X should be relted to y without m in the model. Should be a reduction in significance levels of x once include m.
